The fight against greenwashing starts with AI. Here’s why
July 2023 was the hottest month on record. Around the same time that viral photos showed wildfire smoke smothering Manhattan skyscrapers, investment experts decried ESG metrics as “oftentimes…very subjective, fluffy and easily gamed.” It’s no wonder that accusations of greenwashing levied at industries from oil and gas to finance and banking continue to dominate headlines and boardroom discussions.
At one end of the spectrum, there is a desire to eliminate investing based on environmental, social, and governance (ESG) principles. At the other end, there are urgent calls for the sustainability movement to speed up. There’s also a middle ground everyone seems to agree on: current methods of measuring the effectiveness of ESG efforts are often inconsistent and unreliable.
The transition toward a net-zero future is very much on the horizon. Today, hundreds of firms representing $130 trillion of assets under management are now part of the Glasgow Financial Alliance for Net Zero (GFANZ) and are committed to accelerating the decarbonization of the economy. With the financial services industry heavily involved in driving efforts toward the energy transition, it needs to rethink how to build trust by navigating these ESG measurement challenges. Something needs to be done–and fast.
Making ESG metrics more meaningful
At present, there are dozens of third-party ESG data and ratings providers, among them ESG Book, Moody’s, S&P, Bloomberg, MSCI, Refinitiv, and Sustainalytics. Hundreds of data points are crunched and corralled into weighted environmental, social, and governance categories to arrive at a rating. All that data can be sliced and diced into myriad dimensions, making it hard to differentiate fact from fiction.
For instance, what happens when you look at the overall ESG ratings for a large oil and gas company from three different providers? They span a spectrum: Provider 1 rates it as “BBB” (average), Provider 2 as “good,” and Provider 3 as “severe risk”.
Financial services firms are in a prime position to bring clarity in the current scenario. In a 2022 Dow Jones survey of financial leaders, two-thirds of respondents said that ESG investing is a top driver of sustained, long-term growth. Unsurprisingly, 52% said that the quality of today’s ESG data is not yet good enough to support investment decisions and 58% said greater transparency in how ESG ratings are developed is needed.
Lawsuits. Administrative fines. Reputational damage. The fallout from litigation related to climate change is as real as rising ocean levels. The number of litigation cases currently totals over 2,000, more than a quarter of which were filed between 2020 and 2022. Fitch Ratings said that in Europe, “Climate-related lawsuits targeting banks could set precedents and spur banks to accelerate their carbon-neutrality strategies and their phase-out of fossil fuel financing.”
The stakes to shore up ESG ratings have never been higher across all sectors.
Enter artificial intelligence
Ambiguity around ESG ratings exists for a number of reasons. Siloed and overlapping data, incomplete and exaggerated self-reporting, the lack of measurement standards, and bona fide greenwashing are all culprits.
ESG scores are aggregated. And although there’s some opportunity to drill down, decision-makers require more granular data and transparency to contextualize and determine the reliability and validity of their business decisions, especially when there are outliers. For example, earlier this year, a tobacco company was reported to have a significantly higher ESG score than electric vehicle manufacturer Tesla.
When looking at ESG data, research analysts, underwriters, and asset managers should be able to click through to the source to evaluate it and make comparisons to other comparable ratings and research. And therein lies another issue: ESG scores are partly based on publicly available information, so they depend on a company’s commitment to disclosure and transparency. If gaps exist, how are they filled? It’s hard to say. ESG raters are human, after all. One person’s neutral score is another’s negative, and despite our best efforts, we are all prone to bias and error. Thankfully, artificial intelligence (AI) isn’t sentient.
While ESG ratings are composed of structured, or quantitative data, online chatter and sentiments are unstructured, or qualitative data. Think of the 24-hour news cycle, constantly churning out stories and updates. Unlike a human, machines can read every single article and then analyze text, audio, video, and images. Now consider the dynamic nature of social media platforms, where instant reactions can go viral and affect how stakeholders perceive a company’s ESG practices. Natural language processing (NLP) can “read” human language, and machine learning (ML) can analyze and measure the emotional value or sentiment of communications as positive, neutral, or negative. Analyzing this unstructured data, or sentiment, can help validate and augment ESG claims, uncover discrepancies, and provide real-time information to support investment and underwriting decisions at scale.
Cutting through noise and closing geographic gaps
The unstructured data produced by ESG-related events can often be noisy. Augmenting the output of an NLP sentiment engine with Generative AI for analysis can introduce nuance and precision to both positive and negative sentiments. This approach also helps mitigate any bias in news reports, while advanced filtering further reduces unwanted noise.
For instance, an investment bank was able to create a near real-time ESG sentiment system that integrates more than 70,000 global news sources, enabling traders to quickly understand and react to the impact of events on the ESG profile of companies.
Unstructured data can also help fill in gaps where structured data isn’t available or accessible from third-party raters. In both scenarios, financial institutions can analyze unstructured information to develop proxy data that paints a picture of ESG activities. This information can also aid in creating benchmarks of reports against accepted industry frameworks to reveal inconsistencies and exaggerations in ESG data.
The path forward is simple: To combat greenwashing, the financial services sector needs to develop deeper and more reliable insights into ESG performance. This will ensure they are better placed to take a more predictive look at risks and opportunities. Once we accomplish this, not only can we capitalize on the transition toward a low-carbon economy but we can also gain confidence that the $130 trillion needed to finance it is going into the right hands.
Anirban Bose is CEO of Financial Services and Chairman of APAC at Capgemini.
More must-read commentary published by Fortune:
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.